Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7fad87db0b00>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 100

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7fad8555bcc0>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.2.0
Default GPU Device: /gpu:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    input_real = tf.placeholder(tf.float32,shape=(None,image_width,image_height,image_channels), name ="input_real")
    input_z = tf.placeholder(tf.float32,shape=(None,z_dim), name ="input_z")
    learning_rate = tf.placeholder(tf.float32,name="learning_rate")

    return input_real, input_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

ERROR:tensorflow:==================================
Object was never used (type <class 'tensorflow.python.framework.ops.Operation'>):
<tf.Operation 'assert_rank_2/Assert/Assert' type=Assert>
If you want to mark it as used call its "mark_used()" method.
It was originally created here:
['File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/runpy.py", line 193, in _run_module_as_main\n    "__main__", mod_spec)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/runpy.py", line 85, in _run_code\n    exec(code, run_globals)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel_launcher.py", line 16, in <module>\n    app.launch_new_instance()', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/traitlets/config/application.py", line 658, in launch_instance\n    app.start()', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/kernelapp.py", line 477, in start\n    ioloop.IOLoop.instance().start()', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/zmq/eventloop/ioloop.py", line 177, in start\n    super(ZMQIOLoop, self).start()', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tornado/ioloop.py", line 888, in start\n    handler_func(fd_obj, events)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events\n    self._handle_recv()', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv\n    self._run_callback(callback, msg)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback\n    callback(*args, **kwargs)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 283, in dispatcher\n    return self.dispatch_shell(stream, msg)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 235, in dispatch_shell\n    handler(stream, idents, msg)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 399, in execute_request\n    user_expressions, allow_stdin)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/ipkernel.py", line 196, in do_execute\n    res = shell.run_cell(code, store_history=store_history, silent=silent)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/ipykernel/zmqshell.py", line 533, in run_cell\n    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2698, in run_cell\n    interactivity=interactivity, compiler=compiler, result=result)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2808, in run_ast_nodes\n    if self.run_code(code, result):', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2862, in run_code\n    exec(code_obj, self.user_global_ns, self.user_ns)', 'File "<ipython-input-5-fdf20db26000>", line 23, in <module>\n    tests.test_model_inputs(model_inputs)', 'File "/home/carnd/deep-learning/face_generation/problem_unittests.py", line 12, in func_wrapper\n    result = func(*args)', 'File "/home/carnd/deep-learning/face_generation/problem_unittests.py", line 68, in test_model_inputs\n    _check_input(learn_rate, [], \'Learning Rate\')', 'File "/home/carnd/deep-learning/face_generation/problem_unittests.py", line 34, in _check_input\n    _assert_tensor_shape(tensor, shape, \'Real Input\')', 'File "/home/carnd/deep-learning/face_generation/problem_unittests.py", line 20, in _assert_tensor_shape\n    assert tf.assert_rank(tensor, len(shape), message=\'{} has wrong rank\'.format(display_name))', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tensorflow/python/ops/check_ops.py", line 617, in assert_rank\n    dynamic_condition, data, summarize)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tensorflow/python/ops/check_ops.py", line 571, in _assert_rank_condition\n    return control_flow_ops.Assert(condition, data, summarize=summarize)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 170, in wrapped\n    return _add_should_use_warning(fn(*args, **kwargs))', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 139, in _add_should_use_warning\n    wrapped = TFShouldUseWarningWrapper(x)', 'File "/home/carnd/anaconda3/envs/dl5/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 96, in __init__\n    stack = [s.strip() for s in traceback.format_stack()]']
==================================
Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    with tf.variable_scope('discriminator',reuse = reuse):
        #Input layer is 28*28*3
        x1 = tf.layers.conv2d(images,64,5,strides=2,padding='same')
        relu1 = tf.maximum(0.2*x1,x1)
        #14*14*64
        
        x2 = tf.layers.conv2d(relu1,128,5,strides=2,padding='same')
        bn2 = tf.layers.batch_normalization(x2,training=True)
        relu2 = tf.maximum(0.2*x2,bn2)
        #7*7*128
        
        flat = tf.reshape(relu2,(-1,7*7*128))
        logits = tf.layers.dense(flat,1)
        out = tf.sigmoid(logits)

    return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function
    with tf.variable_scope('generator',reuse=not is_train):
        alpha = 0.2
        x1 = tf.layers.dense(z,7*7*256)
        #reshape to start the convolutional stack
        x1 = tf.reshape(x1,(-1,7,7,256))
        x1 = tf.layers.batch_normalization(x1, training=is_train)
        x1 = tf.maximum(alpha*x1,x1)
        #7*7*256 now
        print ('x1')
        print (x1.shape)
        
        x2 = tf.layers.conv2d_transpose(x1, 128, 5 ,strides=2,padding='same')
        x2 = tf.layers.batch_normalization(x2,training=is_train)
        x2 = tf.maximum(alpha*x2,x2)
        #14*14*128 now
        print ('x2')
        print (x2.shape)
        
        logits = tf.layers.conv2d_transpose(x2, out_channel_dim ,5,strides=2,padding='same')
        #28*28*3 now
        print ('logits')
        print (logits.shape)
        
        out  = tf.tanh(logits)
           
    return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 5)
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 5)
Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    g_model = generator(input_z,out_channel_dim)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse = True)
    #nm can be improved
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)) )
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake
        
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 4)
Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    
    #Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate,beta1=beta1).minimize(d_loss,var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate,beta1=beta1).minimize(g_loss,var_list=g_vars)
    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    _, image_width, image_height, image_channels = data_shape
    input_real, input_z, lr = model_inputs(image_width, image_height, image_channels, z_dim)
    d_loss, g_loss = model_loss(input_real, input_z, image_channels)
    d_opt,g_opt = model_opt(d_loss, g_loss, learning_rate, beta1)
    
    steps = 0
    print_every = 10
    show_every = 100
    losses = []
    n_images = 25 
   
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                # TODO: Train Model
                steps = steps+1
                batch_images = batch_images * 2.0                
                             
                # Random Noise for G
                batch_z = np.random.uniform(-1,1,size=(batch_size,z_dim))
                
                # Run Optimizers
                _ = sess.run(d_opt, feed_dict = {input_real:batch_images, input_z:batch_z, lr:learning_rate})
                _ = sess.run(g_opt, feed_dict = {input_real:batch_images,input_z:batch_z, lr:learning_rate})
                
                if steps % print_every ==0:
                    train_loss_d = d_loss.eval({input_z:batch_z, input_real:batch_images})
                    train_loss_g = g_loss.eval({input_z:batch_z})
                    
                    print("Epoch {}/{}..." .format(epoch_i+1,epoch_count),
                    "Discriminator Loss: {:.4f}..." .format(train_loss_d),
                    "Generator Loss: {:.4f}" .format(train_loss_g))
                    
                    losses.append((train_loss_d,train_loss_g))
                
                if steps % show_every == 0:
                    show_generator_output(sess, n_images, input_z, image_channels, data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 128
z_dim = 100
learning_rate = .002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)
Epoch 1/2... Discriminator Loss: 0.0356... Generator Loss: 10.4272
Epoch 1/2... Discriminator Loss: 0.6567... Generator Loss: 2.8124
Epoch 1/2... Discriminator Loss: 1.8992... Generator Loss: 2.5254
Epoch 1/2... Discriminator Loss: 1.3797... Generator Loss: 0.6595
Epoch 1/2... Discriminator Loss: 2.4759... Generator Loss: 0.1190
Epoch 1/2... Discriminator Loss: 1.2261... Generator Loss: 1.1156
Epoch 1/2... Discriminator Loss: 1.2636... Generator Loss: 0.6088
Epoch 1/2... Discriminator Loss: 2.2931... Generator Loss: 2.3871
Epoch 1/2... Discriminator Loss: 1.8510... Generator Loss: 1.9600
Epoch 1/2... Discriminator Loss: 1.5822... Generator Loss: 0.3726
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 1/2... Discriminator Loss: 1.4141... Generator Loss: 0.3978
Epoch 1/2... Discriminator Loss: 1.2310... Generator Loss: 1.1833
Epoch 1/2... Discriminator Loss: 1.4502... Generator Loss: 0.4461
Epoch 1/2... Discriminator Loss: 1.2733... Generator Loss: 0.8537
Epoch 1/2... Discriminator Loss: 1.3233... Generator Loss: 1.5880
Epoch 1/2... Discriminator Loss: 1.5749... Generator Loss: 2.1509
Epoch 1/2... Discriminator Loss: 1.8985... Generator Loss: 0.2338
Epoch 1/2... Discriminator Loss: 1.3936... Generator Loss: 0.4442
Epoch 1/2... Discriminator Loss: 1.1759... Generator Loss: 1.0393
Epoch 1/2... Discriminator Loss: 1.5972... Generator Loss: 0.3231
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 1/2... Discriminator Loss: 1.1316... Generator Loss: 0.5845
Epoch 1/2... Discriminator Loss: 1.1548... Generator Loss: 0.5335
Epoch 1/2... Discriminator Loss: 1.2123... Generator Loss: 1.4103
Epoch 1/2... Discriminator Loss: 1.1415... Generator Loss: 0.6756
Epoch 1/2... Discriminator Loss: 1.4043... Generator Loss: 0.4112
Epoch 1/2... Discriminator Loss: 1.1353... Generator Loss: 1.1696
Epoch 1/2... Discriminator Loss: 1.0753... Generator Loss: 1.4578
Epoch 1/2... Discriminator Loss: 1.4541... Generator Loss: 1.7301
Epoch 1/2... Discriminator Loss: 1.4603... Generator Loss: 2.1318
Epoch 1/2... Discriminator Loss: 1.1335... Generator Loss: 0.5741
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 1/2... Discriminator Loss: 1.4225... Generator Loss: 0.4240
Epoch 1/2... Discriminator Loss: 1.2248... Generator Loss: 1.6455
Epoch 1/2... Discriminator Loss: 1.0390... Generator Loss: 0.8844
Epoch 1/2... Discriminator Loss: 1.9011... Generator Loss: 0.2140
Epoch 1/2... Discriminator Loss: 1.1271... Generator Loss: 0.6563
Epoch 1/2... Discriminator Loss: 1.1551... Generator Loss: 0.7449
Epoch 1/2... Discriminator Loss: 1.1728... Generator Loss: 1.0087
Epoch 1/2... Discriminator Loss: 1.2407... Generator Loss: 1.2024
Epoch 1/2... Discriminator Loss: 1.3234... Generator Loss: 0.4762
Epoch 1/2... Discriminator Loss: 1.6514... Generator Loss: 1.8780
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 1/2... Discriminator Loss: 1.1557... Generator Loss: 0.9208
Epoch 1/2... Discriminator Loss: 1.2308... Generator Loss: 0.5524
Epoch 1/2... Discriminator Loss: 1.1789... Generator Loss: 0.7515
Epoch 1/2... Discriminator Loss: 1.4637... Generator Loss: 0.3544
Epoch 1/2... Discriminator Loss: 1.4790... Generator Loss: 0.3536
Epoch 1/2... Discriminator Loss: 1.1287... Generator Loss: 0.7711
Epoch 2/2... Discriminator Loss: 1.0119... Generator Loss: 0.9830
Epoch 2/2... Discriminator Loss: 1.2327... Generator Loss: 0.5417
Epoch 2/2... Discriminator Loss: 1.1964... Generator Loss: 1.1312
Epoch 2/2... Discriminator Loss: 1.1409... Generator Loss: 0.7380
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 2/2... Discriminator Loss: 1.4260... Generator Loss: 0.4120
Epoch 2/2... Discriminator Loss: 1.3513... Generator Loss: 0.4231
Epoch 2/2... Discriminator Loss: 1.8817... Generator Loss: 0.2171
Epoch 2/2... Discriminator Loss: 1.1314... Generator Loss: 0.8081
Epoch 2/2... Discriminator Loss: 1.1622... Generator Loss: 0.5612
Epoch 2/2... Discriminator Loss: 1.3211... Generator Loss: 1.6550
Epoch 2/2... Discriminator Loss: 1.6210... Generator Loss: 0.3020
Epoch 2/2... Discriminator Loss: 1.3777... Generator Loss: 0.4137
Epoch 2/2... Discriminator Loss: 1.3266... Generator Loss: 0.4413
Epoch 2/2... Discriminator Loss: 1.1682... Generator Loss: 0.6521
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 2/2... Discriminator Loss: 1.1404... Generator Loss: 0.6023
Epoch 2/2... Discriminator Loss: 1.3686... Generator Loss: 0.4354
Epoch 2/2... Discriminator Loss: 1.0995... Generator Loss: 0.7232
Epoch 2/2... Discriminator Loss: 1.0457... Generator Loss: 0.8213
Epoch 2/2... Discriminator Loss: 0.9937... Generator Loss: 1.0738
Epoch 2/2... Discriminator Loss: 1.5343... Generator Loss: 0.3401
Epoch 2/2... Discriminator Loss: 1.3985... Generator Loss: 1.7590
Epoch 2/2... Discriminator Loss: 1.0253... Generator Loss: 0.8346
Epoch 2/2... Discriminator Loss: 1.0490... Generator Loss: 0.8146
Epoch 2/2... Discriminator Loss: 1.0269... Generator Loss: 1.0882
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 2/2... Discriminator Loss: 1.0482... Generator Loss: 1.4692
Epoch 2/2... Discriminator Loss: 1.5233... Generator Loss: 0.3531
Epoch 2/2... Discriminator Loss: 1.3089... Generator Loss: 1.7439
Epoch 2/2... Discriminator Loss: 1.4003... Generator Loss: 0.3806
Epoch 2/2... Discriminator Loss: 1.2155... Generator Loss: 0.5885
Epoch 2/2... Discriminator Loss: 1.2234... Generator Loss: 1.6425
Epoch 2/2... Discriminator Loss: 1.1086... Generator Loss: 0.6152
Epoch 2/2... Discriminator Loss: 1.3489... Generator Loss: 1.9689
Epoch 2/2... Discriminator Loss: 1.0492... Generator Loss: 1.2438
Epoch 2/2... Discriminator Loss: 1.2396... Generator Loss: 0.5451
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 2/2... Discriminator Loss: 1.0036... Generator Loss: 1.2299
Epoch 2/2... Discriminator Loss: 1.3481... Generator Loss: 0.4207
Epoch 2/2... Discriminator Loss: 1.0162... Generator Loss: 1.0417
Epoch 2/2... Discriminator Loss: 1.5496... Generator Loss: 2.3555
Epoch 2/2... Discriminator Loss: 1.5231... Generator Loss: 0.3492
Epoch 2/2... Discriminator Loss: 1.2152... Generator Loss: 0.5347
Epoch 2/2... Discriminator Loss: 1.6151... Generator Loss: 2.3550
Epoch 2/2... Discriminator Loss: 1.2274... Generator Loss: 0.5073
Epoch 2/2... Discriminator Loss: 0.9303... Generator Loss: 1.1985
Epoch 2/2... Discriminator Loss: 1.2254... Generator Loss: 1.6091
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 1)

Epoch 2/2... Discriminator Loss: 1.0535... Generator Loss: 1.2835
Epoch 2/2... Discriminator Loss: 1.4025... Generator Loss: 2.0007
Epoch 2/2... Discriminator Loss: 1.0118... Generator Loss: 1.0463

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 64
z_dim = 100
learning_rate = 0.002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 5

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)
Epoch 1/5... Discriminator Loss: 1.2120... Generator Loss: 1.1526
Epoch 1/5... Discriminator Loss: 0.0481... Generator Loss: 4.4557
Epoch 1/5... Discriminator Loss: 0.0849... Generator Loss: 5.6036
Epoch 1/5... Discriminator Loss: 2.2110... Generator Loss: 0.1904
Epoch 1/5... Discriminator Loss: 0.5816... Generator Loss: 4.0569
Epoch 1/5... Discriminator Loss: 0.8203... Generator Loss: 2.1369
Epoch 1/5... Discriminator Loss: 0.3128... Generator Loss: 2.3517
Epoch 1/5... Discriminator Loss: 0.7053... Generator Loss: 3.4676
Epoch 1/5... Discriminator Loss: 1.9197... Generator Loss: 0.3427
Epoch 1/5... Discriminator Loss: 1.2297... Generator Loss: 0.6150
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.7293... Generator Loss: 0.3377
Epoch 1/5... Discriminator Loss: 1.3081... Generator Loss: 2.8253
Epoch 1/5... Discriminator Loss: 1.4452... Generator Loss: 0.4009
Epoch 1/5... Discriminator Loss: 1.5567... Generator Loss: 2.4701
Epoch 1/5... Discriminator Loss: 1.1406... Generator Loss: 2.5245
Epoch 1/5... Discriminator Loss: 1.4562... Generator Loss: 0.9402
Epoch 1/5... Discriminator Loss: 0.9217... Generator Loss: 1.2516
Epoch 1/5... Discriminator Loss: 1.7130... Generator Loss: 1.6319
Epoch 1/5... Discriminator Loss: 1.4773... Generator Loss: 2.5195
Epoch 1/5... Discriminator Loss: 0.8042... Generator Loss: 0.9859
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.1259... Generator Loss: 1.2994
Epoch 1/5... Discriminator Loss: 1.4154... Generator Loss: 0.5938
Epoch 1/5... Discriminator Loss: 1.3875... Generator Loss: 2.6220
Epoch 1/5... Discriminator Loss: 1.5017... Generator Loss: 0.8181
Epoch 1/5... Discriminator Loss: 1.1092... Generator Loss: 1.5520
Epoch 1/5... Discriminator Loss: 1.5749... Generator Loss: 2.3384
Epoch 1/5... Discriminator Loss: 1.1195... Generator Loss: 0.5756
Epoch 1/5... Discriminator Loss: 1.2656... Generator Loss: 0.6797
Epoch 1/5... Discriminator Loss: 1.7918... Generator Loss: 0.2494
Epoch 1/5... Discriminator Loss: 1.1177... Generator Loss: 0.9150
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.2268... Generator Loss: 1.5043
Epoch 1/5... Discriminator Loss: 1.2392... Generator Loss: 0.5561
Epoch 1/5... Discriminator Loss: 1.7011... Generator Loss: 1.8755
Epoch 1/5... Discriminator Loss: 1.5162... Generator Loss: 3.5147
Epoch 1/5... Discriminator Loss: 2.5934... Generator Loss: 2.9261
Epoch 1/5... Discriminator Loss: 1.0734... Generator Loss: 0.6230
Epoch 1/5... Discriminator Loss: 1.2034... Generator Loss: 0.4712
Epoch 1/5... Discriminator Loss: 0.9182... Generator Loss: 2.7681
Epoch 1/5... Discriminator Loss: 1.7365... Generator Loss: 2.7420
Epoch 1/5... Discriminator Loss: 0.9183... Generator Loss: 0.8731
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.1251... Generator Loss: 0.7126
Epoch 1/5... Discriminator Loss: 1.5665... Generator Loss: 0.3191
Epoch 1/5... Discriminator Loss: 1.5425... Generator Loss: 2.4444
Epoch 1/5... Discriminator Loss: 1.4933... Generator Loss: 0.4053
Epoch 1/5... Discriminator Loss: 0.7212... Generator Loss: 1.5408
Epoch 1/5... Discriminator Loss: 1.4759... Generator Loss: 0.4414
Epoch 1/5... Discriminator Loss: 2.0181... Generator Loss: 0.1950
Epoch 1/5... Discriminator Loss: 0.9929... Generator Loss: 0.6578
Epoch 1/5... Discriminator Loss: 1.3512... Generator Loss: 0.3967
Epoch 1/5... Discriminator Loss: 1.3531... Generator Loss: 0.4101
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.3492... Generator Loss: 0.8942
Epoch 1/5... Discriminator Loss: 1.4776... Generator Loss: 0.4121
Epoch 1/5... Discriminator Loss: 1.5684... Generator Loss: 2.6873
Epoch 1/5... Discriminator Loss: 1.5911... Generator Loss: 0.3037
Epoch 1/5... Discriminator Loss: 1.1543... Generator Loss: 2.1841
Epoch 1/5... Discriminator Loss: 1.0535... Generator Loss: 0.7481
Epoch 1/5... Discriminator Loss: 0.5336... Generator Loss: 1.8824
Epoch 1/5... Discriminator Loss: 2.4134... Generator Loss: 0.1222
Epoch 1/5... Discriminator Loss: 1.3395... Generator Loss: 1.6369
Epoch 1/5... Discriminator Loss: 0.8859... Generator Loss: 1.7169
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 0.9271... Generator Loss: 1.3111
Epoch 1/5... Discriminator Loss: 0.9847... Generator Loss: 1.2012
Epoch 1/5... Discriminator Loss: 1.1357... Generator Loss: 0.9972
Epoch 1/5... Discriminator Loss: 0.9558... Generator Loss: 1.0042
Epoch 1/5... Discriminator Loss: 0.6025... Generator Loss: 1.6658
Epoch 1/5... Discriminator Loss: 1.4083... Generator Loss: 0.4479
Epoch 1/5... Discriminator Loss: 1.5829... Generator Loss: 2.6158
Epoch 1/5... Discriminator Loss: 0.9787... Generator Loss: 2.2423
Epoch 1/5... Discriminator Loss: 1.0697... Generator Loss: 1.1132
Epoch 1/5... Discriminator Loss: 1.1704... Generator Loss: 0.5625
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.2935... Generator Loss: 2.1010
Epoch 1/5... Discriminator Loss: 0.8902... Generator Loss: 1.9325
Epoch 1/5... Discriminator Loss: 0.9962... Generator Loss: 1.0060
Epoch 1/5... Discriminator Loss: 1.5895... Generator Loss: 0.2878
Epoch 1/5... Discriminator Loss: 1.3510... Generator Loss: 1.4281
Epoch 1/5... Discriminator Loss: 0.9831... Generator Loss: 0.8496
Epoch 1/5... Discriminator Loss: 1.2728... Generator Loss: 3.0288
Epoch 1/5... Discriminator Loss: 1.7336... Generator Loss: 0.3684
Epoch 1/5... Discriminator Loss: 0.7864... Generator Loss: 0.9755
Epoch 1/5... Discriminator Loss: 0.3294... Generator Loss: 2.4198
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 2.8199... Generator Loss: 4.1686
Epoch 1/5... Discriminator Loss: 1.1534... Generator Loss: 3.4907
Epoch 1/5... Discriminator Loss: 0.6662... Generator Loss: 2.2113
Epoch 1/5... Discriminator Loss: 0.9289... Generator Loss: 0.7878
Epoch 1/5... Discriminator Loss: 0.9824... Generator Loss: 1.4561
Epoch 1/5... Discriminator Loss: 0.5535... Generator Loss: 1.8862
Epoch 1/5... Discriminator Loss: 0.7992... Generator Loss: 1.6071
Epoch 1/5... Discriminator Loss: 0.5139... Generator Loss: 2.3492
Epoch 1/5... Discriminator Loss: 0.6282... Generator Loss: 1.6527
Epoch 1/5... Discriminator Loss: 0.9293... Generator Loss: 1.0867
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Epoch 1/5... Discriminator Loss: 1.2467... Generator Loss: 0.8509
Epoch 1/5... Discriminator Loss: 1.4346... Generator Loss: 0.4181
Epoch 1/5... Discriminator Loss: 1.0737... Generator Loss: 0.8510
Epoch 1/5... Discriminator Loss: 0.9684... Generator Loss: 1.0832
Epoch 1/5... Discriminator Loss: 1.5005... Generator Loss: 0.3829
Epoch 1/5... Discriminator Loss: 1.6527... Generator Loss: 0.3041
Epoch 1/5... Discriminator Loss: 1.0540... Generator Loss: 2.7848
Epoch 1/5... Discriminator Loss: 0.7195... Generator Loss: 1.9736
Epoch 1/5... Discriminator Loss: 1.3829... Generator Loss: 0.4709
Epoch 1/5... Discriminator Loss: 0.9851... Generator Loss: 1.3498
x1
(?, 7, 7, 256)
x2
(?, 14, 14, 128)
logits
(?, 28, 28, 3)

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.